Hyperspectral images offer rich spectral information but suffer from high dimensionality and redundancy, increasing the computational burden for classification. This paper presents an unsupervised band selection method that leverages Kernel Fuzzy C-Means clustering, guided by a Normalized Mutual Information based dissimilarity measure, to group spectrally similar bands. To ensure robustness against clustering variability, the process is repeated over multiple simulations, and representative medoid bands are consistently selected from each cluster. Finally, these selected bands are ranked based on their selection frequency and average membership strength. Experiments on the Indian Pines and Kennedy Space Center datasets demonstrate the superior performance of the proposed method over several state-of-the-art algorithms.

High spectral dimension of the hyperspectral images renders classification task computationally intensive. To mitigate this, a fully unsupervised band selection and ranking algorithm is proposed that leverages K-medoids clustering and mutual information to identify a subset of informative, uncorrelated spectral bands. The proposed method’s performance is assessed on the Botswana and Indian Pines datasets, outperforming several other mutual information based approaches with respect to overall classification accuracy. These results validate the efficacy of the proposed method for hyperspectral data analysis.

Purpose: Health recommenders act as important decision support systems, aiding patients and medical professionals in taking actions that lead to patients' well-being. These systems extract the information which may be of particular relevance to the end-user, helping them in making appropriate decisions. The present study proposes a feature recommender, as a part of a disease management system, that identifies and recommends the most important risk factors for an illness. Methods: A novel mutual information and ensemble-based feature ranking approach for identifying critical risk factors in healthcare prognosis is proposed. Results: To establish the effectiveness of the proposed method, experiments have been conducted on four benchmark datasets of diverse diseases (clear cell renal cell carcinoma (ccRCC), chronic kidney disease, Indian liver patient, and cervical cancer risk factors). The performance of the proposed recommender is compared with four state-of-the-art methods using recommender systems' performance metrics like average precision@K, precision@K, recall@K, F1@K, reciprocal rank@K. The method is able to recommend all relevant critical risk factors for ccRCC. It also attains a higher accuracy (96.6 support vector machine and neural network, respectively) for ccRCC staging with a reduced feature set as compared to existing methods. Moreover, the top two features recommended using the proposed method with ccRCC, viz. size of tumor and metastasis status, are medically validated from the existing TNM system. Results are also found to be superior for the other three datasets. Conclusion: The proposed recommender can identify and recommend risk factors that have the most discriminating power for detecting diseases.

In this paper, we address the problem of smoke plume segmentation from background clutter. Smoke plumes can be generated from fires, explosions, etc. In the mining industry, plumes from blasts need to be characterized in terms of their volume and concentration, for example. Plume segmentation is required in order to start such an analysis. We present a new image processing approach based on a fast local Laplacian filtering (FLLF) technique. In addition, we discuss how we designed and executed our own field experiments to acquire actual test data of smoke plumes from RGB video cameras. Lastly, we show how the FLLF technique can be used to generate thousands of training samples with applications in machine learning. Results show that the FLLF technique outperforms state-of-the-art approaches (i.e., SFFCM and an approach by Wang et al.) when tested using metrics such as Accuracy, the Jaccard Index, F1-score, False Alarms and Misses. We also show that the FLLF technique is more computationally efficient.

Classification of hyperspectral images is computationally expensive due to the presence of large number of spectral bands. Therefore, dimensionality reduction using selection of optimal set of bands is an essential task to speed up the subsequent classification process. Bands must be selected in such a way so that they are as much independent as possible without sacrificing classification accuracy. In this context, a supervised band selection approach is proposed combining mutual and neighborhood information of bands. For classification purpose, Support Vector Machine classifier is used. Overall classification accuracy is considered to assess the efficiency of the proposed method. Performance of the proposed technique is compared with several other Mutual Information based methods and the proposed method is found to be better as compared to others.

Kidney is an essential organ in human body. It maintains homeostasis and removes harmful substances through urine. Renal cell carcinoma (RCC) is the most common form of kidney cancer. Around 90\% of all kidney cancers are attributed to RCC. Most harmful type of RCC is clear cell renal cell carcinoma (ccRCC) that makes up about 80\% of all RCC cases. Early and accurate detection of ccRCC is necessary to prevent further spreading of the disease in other organs. In this article, a detailed experimentation is done to identify important features which can aid in diagnosing ccRCC at different stages. The ccRCC dataset is obtained from The Cancer Genome Atlas (TCGA). A novel mutual information and ensemble based feature ranking approach considering the order of features obtained from 8 popular feature selection methods is proposed. Performance of the proposed method is evaluated by overall classification accuracy obtained using 2 different classifiers (ANN and SVM). Experimental results show that the proposed feature ranking method is able to attain a higher accuracy (96.6\% and 98.6\% using SVM and NN, respectively) for classifying different stages of ccRCC with a reduced feature set as compared to existing work. It is also to be noted that, out of 3 distinguishing features as mentioned by the existing TNM system (proposed by AJCC and UICC), our proposed method was able to select two of them (size of tumour, metastasis status) as the top-most ones. This establishes the efficacy of our proposed approach.

Hyperspectral sensors obtain a set of images from hundreds of contiguous and narrow bands of electromagnetic spectrum from visible to infrared regions. As large number of bands are present in these images, computational overhead for classifying them is very high. To speed up classification, reducing dimensionality by proper selection of a subset of bands is necessary. A filter based method using mutual information is proposed in this regard. Classification is done using Support Vector Machine classifier. Two evaluation measures, namely, overall classification accuracy and Kappa coefficient are considered to assess the efficiency of the proposed method. Performance of the proposed technique is compared with two other mutual information based methods and the proposed method is found to be better as compared to others.

The COVID-19 pandemic has affected humans worldwide, and we are in dire need of techniques to bring this situation within our control. Among the various approaches attempted by researchers, preliminary prediction of COVID-19 through chest X-ray images is proving to be quite beneficial and thus, is being explored thoroughly. In this paper, a novel combination of local binary pattern based feature selection along with a convolutional neural network is proposed which can predict positive and negative cases by analysing chest X-ray images. The model consists of a feature extraction process followed by various pooling and convolution layers systematically placed to give an optimal output. The proposed model has been trained and tested on a COVID-19 CXR images dataset, and it is seen that it achieves a significant improvement over the five other comparison methods.

Disease detection in crops and plants is essential for production of good and improved quality of food, life and a stable agricultural economy. It becomes tedious and time consuming to observe the infected parts of plants manually. Recourses with proper expertise are also required to have continuous monitoring. Digital image processing along with computer vision techniques can be applied to automate early detection of plant diseases and it can save significant amount of resources. In this paper, an automated approach based on image processing and machine learning techniques is proposed which can detect three major kinds of diseases Downy Mildew, Frogeye Leaf Spot and Septoria Leaf Blight that affects apple, grapes, soybean, tomatoes and many other major plants of economic value. Generally, leaves are the most affected part of the plants. So, instead of the whole plant, concentration is given on the leaf. In this paper, image pre-processing methods like noise removal and contrast enhancement followed by colour space transformation and k-means clustering is used to segment affected parts of soybean leaves, after that both texture and colour features are extracted from segmented samples and Support Vector Machine (SVM) classification is used to separate three kinds of diseases mentioned previously.

Plant diseases have turned into a problem as it can cause substantial decrease in both quality and quantity of agricultural harvests. Image processing can help in the following issues: early detection which leads to better growth of plant, and suggestion of the type and amount of pesticides knowing the pest. Leaves and stems are the most affected part of the plants. So, they are the study of interest. The beetle can affect the leaves, which leads to severe harm in the plant. In this paper, we propose an automated method for classification of various types of beetles, which consists of (i) image preprocessing techniques, such as contrast enhancement, are used to improve the quality of image which makes advance processing satisfactory; (ii) K-means clustering method is applied for segmenting pest from infected leaves; (iii) 24 features are extracted from those segmented images by using feature extraction, mainly GLCM; and (iv) support vector machine is used for multi-classification of the beetles. The proposed algorithm can successfully detect and classify the 12 classes of beetles with an accuracy of 89.17%, which outperforms the other multi-class pest-classification algorithm by a decent margin.

Glaucoma is one of the eye diseases that can lead to the blindness if not detected and treated at proper time. This paper presents a novel technique to diagnose glaucoma using digital fundus images. In this proposed method, the objective is to apply image processing and machine-learning techniques on the digital fundus images of the eye for separating glaucomatous eye from normal eye. Image preprocessing, techniques such as noise removal and contrast enhancement are used for improving the quality of image thus making it suitable for further processing. Statistical feature extraction methods such as Gray-Level Run Length Matrix (GLRLM) and Gray-Level Co-occurrence Matrix (GLCM) are used for extracting texture features from preprocessed fundus images. Support Vector Machine (SVM) classification method is used for distinguishing glaucomatous eye fundus images from normal, unaffected eye fundus images. The performance of the trained SVM classifier is also tested on a test set of eye fundus images and comparison is done with other existing recent methods of Glaucoma detection.

---------------------------------------------------------------------***--------------------------------------------------------------------Abstract Pest detection in plants and crops is essential for production of good quality food, improved quality of life and a stable agricultural economy. Excessive use of pesticides for pest control is harmful to plants, animals as well as human beings. Digital image processing along with computer vision techniques can be applied for early detection of pests and it can minimize amount of pesticides used in the plants. Generally, leaves are the most affected part of the plants. So, the study of interest is the leaf, rather than whole plant. Among many pests, the white fly is one of the most hazardous pests that affect the leaves. This paper presents an automated approach for detection of white fly pest from leaf images of various plants. Initially, image pre-processing techniques such as noise removal and contrast enhancement are used for improving the quality of image thus making it suitable for further processing. Then, k-means clustering method is used for segmenting pest from infected leaves. After that, texture features are extracted from those segmented images by statistical feature extraction methods such as Gray Level Run Length Matrix (GLRLM) and Gray Level Co-occurrence Matrix (GLCM). Finally, various classifiers like Support Vector machine, Artificial Neural Network, Bayesian classifier, Binary decision tree classifier and k-Nearest neighbor classifier are used to distinguish between healthy leaf images from white fly pest infected leaf images.

Remote sensing image fusion is a process that integrates the spatial detail of panchromatic (PAN) image and the spectral information of a low-resolution multispectral (MS) image and produces a fused image that contain both high spatial and spectral details. In this paper, a new remote sensing image fusion method is proposed based on Statistical Univariate Finite Mixture Model (UFMM) in Shearlet Domain. Foremost, the Shearlet sub-bands for PAN and MS image are achieved by Shearlet Transform (ST). Latter, a novel fusion strategy is designed for both low and high pass sub-bands. Finally, the fused image is achieved by Inverse Shearlet Transform (IST). By comparing with the well-known methods in terms of several quality evaluation indexes, the experimental results on QuickBird and IKONOS images show the superiority of our method.

Preservation of spectral information and enhancement of spatial resolution is the most important issue in remote sensing image fusion. In this paper, a new remote sensing satellite image fusion method using shearlet transform (ST) with Hausdorff fractal dimension(HFD) estimation method is proposed. Firstly, ST is used in each high-spatial-resolution panchromatic (PAN) image and multi-spectral image (MS). Then, the low frequency sub-band coefficients from different images are combined according to the HFD method which estimates and selects the modified low-pass band automatically. The composition of different high-pass sub-band coefficients achieved by the ST decomposition is discussed in detail. Finally, we achieve fusion results from the inverse transformation of ST. Experimental results show that the proposed method outperforms many state-of-the-art techniques in both subjective and objective evaluation measures.